Gastric Cancer Extracellular Vesicles Tune the Migration and Invasion of Epithelial and Mesenchymal Cells in a Histotype-Dependent Manner

Abstract

Extracellular vesicles (EVs) secreted by tumor cells modulate recipient cells' behavior, but their effects in normal cells from the tumor microenvironment remain poorly known. In this study, we dissected the functional impact of gastric cancer cell-derived EVs (GC-EVs), representative of distinct GC histotypes, on the behavior of normal isogenic epithelial and mesenchymal cells. GC-EVs were isolated by differential centrifugation and characterized by transmission electron microscopy, nanoparticle tracking analysis, and imaging flow-cytometry. Epithelial and mesenchymal cells were challenged with GC-EVs and submitted to proliferation, migration, and invasion assays. Expression of epithelial and mesenchymal markers was followed by immunofluorescence and flow-cytometry. Our results indicated that GC-EVs secreted by diffuse-type cancer cells decrease the migration of recipient cells. This effect was more prominent and persistent for mesenchymal recipient cells, which also increased Fibronectin expression in response to EVs. GC-EVs secreted by cancer cells derived from tumors with an intestinal component increased invasion of recipient epithelial cells, without changes in EMT markers. In summary, this study demonstrated that GC-EVs modulate the migration and invasion of epithelial and mesenchymal cells from the tumor microenvironment, in a histotype-dependent manner, highlighting new features of intestinal and diffuse-type GC cells, which may help explaining differential metastasis patterns and aggressiveness of GC histotypes.

Keywords: epithelial-to-mesenchymal transition; extracellular vesicles; gastric cancer; invasion; migration.

Figure 1

Scheme illustrating the experimental design of the study. Four distinct gastric cancer (GC) cell lines (MKN74, MKN45, Kato III, and IPA220) were used as donor cells of extracellular vesicles (EVs), which were isolated by differential centrifugation and characterized by transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), and imaging flow cytometry. Epithelial (red) and mesenchymal (green) cells were used as recipients of GC-EVs and submitted to distinct functional assays.

Figure 2

Characterization of a human Transforming Growth Factor beta TGFβ-induced epithelial to mesenchymal transition (EMT) cell model. (A) Bright-field microscopic images of epithelial and mesenchymal cells; (B) quantification of epithelial (CDH1 and OCLN) and mesenchymal markers (CDH2, VIM, and FN) mRNA expression by qRT-PCR. Data obtained for M cells was normalized for E cells and GAPDH was used as endogenous control. Graphs represent the mean ± standard deviation of three independent experiments (*p ≤ 0.05, unpaired t-test with Welch’s correction); and (C) immunofluorescence for E-cadherin (red staining), Fibronectin (green staining) in epithelial and mesenchymal cells. Scale bar: 20μm. (E—epithelial cells; M—mesenchymal cells).

Figure 3

Characterization of EVs secreted by MKN74, MKN45, Kato III, and IPA220 GC cell lines. (A) Representative electron microscopy images of EVs isolated from GC cells. Scale bars: 200 nm; (B) NTA of isolated EVs with mode size distribution (left) and particle concentration (right). Graphs represent the mean ± standard deviation of at least 14 biological replicates; (C) detection of CD9, CD81, Flotillin-1, and Cytochrome C (negative control) by imaging flow cytometry. Distribution and representative images of the intensity of fluorescence detected for each marker in three biological replicates. Bright-field images (BF) show beads to which EVs were coupled, fluorescence images (AF488) show EVs labeled with specific markers, and merged images (M) show labeled EVs coupled to beads.

Figure 4

GC-EVs do not modulate membrane E-cadherin expression of epithelial and mesenchymal cells. (A) Percentage of E-cadherin-positive cells in non-treated (control) and treated epithelial and mesenchymal cells were assessed by flow cytometry. Graph represents the mean ± standard deviation of three independent experiments (**** p ≤ 0.0001, Two-way ANOVA with Tukey’s multiple comparisons test); (B) intensity of E-cadherin surface expression of non-treated (control) and treated epithelial and mesenchymal cells was assessed by flow cytometry. Graph represents the mean ± standard deviation of three independent experiments, and data were normalized to each respective control cell. TGFβ1 was maintained in M cultures (control and treated M cells).

Figure 5

GC-EVs impair the migration of epithelial and mesenchymal cells. (A) Time-lapse microscopic images illustrating the migration of non-treated (control) and treated epithelial and mesenchymal cells, at distinct time-points; (B,C) the migration rate of non-treated (control) and treated epithelial and mesenchymal cells was assessed by time-lapse microscopy; (B) graphs represent the mean percentage of wound closure ± standard deviation of three independent experiments; (C) graphs represent the mean of normalized wound closure ± standard deviation of three independent experiments; the data was normalized to the control cells at each time-point (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001, and two-way ANOVA with Dunnett’s multiple comparisons test; blue and orange asterisks correspond to E and M cells treated with IPA220-EVs or Kato III-EVs, respectively.

Figure 6

GC-EVs impair the invasion of epithelial and mesenchymal cells. The invasion levels of non-treated (control) and treated epithelial (A) or mesenchymal cells (B) were assessed by matrigel invasion assay; graph represents the mean ± standard deviation of three independent experiments and data was normalized to control cells (*p ≤ 0.05, one-way ANOVA with Kruskal-Wallis test); (C) immunofluorescence for E-cadherin (red staining) and fibronectin (green staining) in non-treated (control) and treated epithelial and mesenchymal cells. Scale bars: 50μm.